Unequal response to mobility restrictions: evidence from COVID-19 lockdown in the city of Bogotá

ABSTRACT

We study the effectiveness of the mobility restrictions imposed by governments to curb urban mobility. We use mobile phone-tracked movements to determine whether users left their homes and explore the role of socio-economic differences across neighbourhoods in explaining their unequal response to lockdown measures. We rely on novel data showing changes in movements in highly disaggregated spatial units in Bogotá, Colombia, before and during the first wave of the COVID-19 pandemic, matched with data on socio-economic characteristics and data on non-pharmaceutical interventions implemented in the period of analysis. We find that the general lockdown imposed in the city significantly reduced mobility (by about 41 percentage points). When looking at the unequal response across locations, we find that low-income areas, with higher population density, informality and overcrowding, reacted less to mobility restrictions.

KEYWORDS:

• mobility
• development
• inequality
• COVID-19
• place-based policies

JEL:

• I18
• R11
• R12

ACKNOWLEDGEMENTS

The authors thank the United Nations Development Programme (UNDP) and GRANDATA for information about mobility provided in the call ‘Exploring Impact and Response to the COVID-19 Pandemic in Latin America and the Caribbean using Mobility Data’, and the UNDP team for all the data help. We also thank the research assistants Ingrid Quevedo, Daniel Jimenez and Carolina Correal who helped to prepare and process the data used for this study.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the author(s).

Notes

1 We focus on mobility change. Ultimately, this will affect public health, as mobility has a strong connection with COVID-19 cases. For instance, Glaeser et al. (2020), using data across five US cities, estimate that total cases per capita decreased by 19% for every 10 percentage points fall in mobility. We also find a positive and significant statistical relationship between mobility and local cases in Bogotá in our sample.

2 For mobility data aggregates, see covid.grandata.com/

3 Geohash is a public domain geocode system invented in 2008 by Gustavo Niemeyer. All locations follow a hierarchical spatial data system dividing space into a grid.

4 Couture et al. (2021) show that smartphone mobility data are representative of movement patterns in the United States similar to conventional survey data.

5 The survey also has information for 17 arbitrary aggregations of UPZ. However, with this information is not possible to recover all UPZs. Furthermore given the sample design of these aggregations, when fractioning them, the representativeness is compromised.

6 We took the average mobility reduction of the five wealthiest and five poorer UPZs for this calculation.

7 Subsidies data come from the city official programme website. For data, see https://rentabasicabogota.gov.co/. The subsidies are from a programme called Solidary Bogotá at Home, which disbursed subsidies to lower income people in Bogotá.

8 The city government provides information in the report at https://Bogotá.gov.co/mi-ciudad/ingreso-de-viajeros-a-colombia

9 Ministry of Health and Social Protection, https://www.ins.gov.co/Noticias/Paginas/Coronavirus.aspx

10 Figures A4 and A6 in Appendix A in the supplemental data online show binned scatterplots. Binned scatterplots are a non-parametric method of plotting the conditional expectation function (which describes the average y-value for each x-value).

11 Additional issues include the unobserved rate of adoption of additional measures, such as masks, that will affect how social contact is reflected in case growth. There is also disagreement about the optimal lag to use to link mobility and cases (estimated between seven and 28 days) (Nie et al., 2020).

12 The full effect of the lockdown on mobility for UPZ $i$ is given by $\mathrm{\delta ln}{M}_{\mathrm{it}}/\delta \mathrm{LockDow}{\mathrm{n}}_t=\widehat{{\text{β}}_i}+\eta$. However, the estimated $\eta$ is a constant across UPZs. ${\text{β}}_i$ captures the heterogeneity of compliance beyond the average for each UPZ, the variation we explain in the second estimation stage through equation (2).

13 $\alpha$ is the average effect of district-specific restrictions. Recall that districts are areas that include multiple UPZs and that all district-specific lockdowns were implemented after the citywide lockdown ended, with no time overlap.

14 Anecdotally, the short notice even caused waves of people to encounter problems when returning to the city from weekend trips.

15 In contrast, district specific lockdowns were implemented according to local COVID cases surges, an endogenous variable to mobility. This makes anticipatory effects likely. We focus on the effect of the main lockdown and use the UPZ-specific lockdown indicators as a control and do not interpret them causally.

16 The two-stage strategy is similar to that followed in other urban economics papers such asCombes et al. (2008) and Roca and Puga (2017), where city productivity is estimated in a first stage and then regressed on explanatory variables in a second stage.

17 Hainmueller et al. (2019) highlight two problems of this type of multiplicative interaction models. First, these models assume a linear interaction effect that changes at a constant rate with the moderator. This is an issue when the treatment variable (lockdown in our analysis) is either binary or continuous and the moderator (UPZ indicator in our analysis) is continuous. When both are binary, as in our case, they suggest using a fully saturated model that dummies out the treatment and the moderator and includes all interaction terms, the approach we follow. Second, estimates of the conditional effects of the independent variable can be misleading if there is a lack of common support of the moderator. This is not an issue, again, because all levels in the interaction terms are parameterised as dummy variables.

18 We estimate independent regressions taking into account that there is a high correlation between several of the socio-economic variables presented in Table 1.

19 The MPI is a comprehensive poverty measure that complements traditional monetary poverty measures with deprivations in health, education, labour market outcomes and living standards.

Additional information

Funding

David Castells-Quintana gratefully acknowledges the support from the Spanish Ministry of Science and Innovation [grant numbers PID 2019-104723RB-100 and PID2020-118800GB-100, and the Catalan Agency for Management of University and Research Grants [grant number 2017SGR1301].